Model-Based Actor-Critic for Multi-Objective Reinforcement Learning with Dynamic Utility Functions.
Johan KällströmFredrik HeintzPublished in: AAMAS (2023)
Keyphrases
- utility function
- actor critic
- reinforcement learning
- multi objective
- model free
- temporal difference
- function approximation
- reinforcement learning algorithms
- policy gradient
- approximate dynamic programming
- optimal control
- decision problems
- evolutionary algorithm
- decision theory
- neuro fuzzy
- decision makers
- gradient method
- optimization algorithm
- markov decision problems
- dynamic environments
- rl algorithms
- state space
- learning algorithm
- probability distribution
- monte carlo
- genetic algorithm
- optimal solution
- policy iteration
- linear program
- reinforcement learning methods
- supervised learning
- step size